Table 1.

Results of the PLS-1 regression models correlating the chemical parameters and the signals of the sensors.

Grape	Parameter	$R_{\text{C}}^2$(a)	RMSEC(b)	$R_{\text{P}}^2$(C)	RMSEP(d)	Number of components
Cabernet	°Brix	0.87	0.86	0.85	0.95	3
	TA (g/L)	0.87	0.81	0.81	1.02	7
	TPI	0.86	2.13	0.83	2.42	2
Garnacha	°Brix	0.92	0.91	0.82	1.40	6
	TA (g/L)	0.85	1.24	0.73	1.72	6
	TPI	0.86	2.25	0.77	3.01	3
Juan Garcia	°Brix	0.86	0.79	0.79	1.00	5
	TA (g/L)	0.80	1.08	0.70	1.39	6
	TPI	0.84	1.54	0.79	1.84	4
Mencia Regadio	°Brix	0.97	0.22	0.96	0.26	5
	TA (g/L)	0.89	0.69	0.86	0.77	5
	TPI	0.88	1.08	0.83	1.33	5
Mencia Secano	°Brix	0.86	0.58	0.79	0.74	6
	TA (g/L)	0.81	0.84	0.71	1.06	6
	TPI	0.86	1.57	0.82	1.79	3
Prieto Picudo	°Brix	0.87	0.73	0.63	1.25	7
	TA (g/L)	0.88	0.93	0.70	1.52	7
	TPI	0.83	1.33	0.81	1.43	3
Rufete	°Brix	0.85	0.49	0.83	0.53	3
	TA (g/L)	0.82	0.91	0.76	1.09	5
	TPI	0.91	1.51	0.90	1.64	3
Tempranillo	°Brix	0.81	0.89	0.76	1.01	6
	TA (g/L)	0.91	0.56	0.76	0.95	6
	TPI	0.87	2.05	0.75	2.94	4

Models were established for each variety of grape considered separately.

(a) Squared correlation coefficient in calibration; (b) Root mean square error of calibration; (c) Squared correlation coefficient in prediction; (d) Root mean square error of prediction.